Fat Interfaces (opposing): The Client Is Too Thin

Fat Interfaces (opposing): The Client Is Too Thin

Post by Kent Paul Dola » Thu, 14 Aug 2003 04:18:27




> Your experience of HTML based interfaces being
> user-hostile is widely spread.

Not among users!  Virtually everyone with a computer is or
wants to be online to use the Internet from a browser.  HTML
interfaces are far and away the most popular interfaces in
the world.  This is absolutely _not_ because users perceive
them as "user-hostile".  Even Internet Explorer, the world's
next biggest software bug farm (behind WinOS collectively),
is enthusiastically tolerated by users because it is so
convenient _despite_ its tendency to crash at the drop of a
mouse click. [Mine froze up writing this posting when I
tried to spawn a new window in which to look up a spelling
at http://www.m-w.com and I had to start again from scratch,
using vim() this time to compose this article as I should
always do for anything over a two-line response.]

Quote:> We have been developing browser based apps for years now
> and have struggled with this (coming from a GUI background
> specializing in [beautiful] user interfaces) for quite a
> long time.

Yet in the event, user's don't _want_ "beautiful"
interfaces, they want _minimalist_ _standard_ _already
learned_ HTML interfaces that add minimal mental effort
into shifting from application to application.  Companies
that pay attention to this trend and port legacy system
interfaces to extremely simple-minded HTML interfaces are
making money hand over fist.  Companies that go on
stubbornly swimming against that tide are very likely to
flounder, fail, and never understand why.

The users' perspective is that however slow the response to
an HTML form submission to the server, the time the user
would have spent doing the mental paradigm shift from one
ideosyncratic interface to another far dominates, both in
real time lost to data entry misunderstandings, and in the
excess mental effort needed compared to shifting seamlessly
between applications with minimalist HTML interfaces.

In today's world, users of robust browsers are free to
multitask among several simultaneous applications or copies
of the same online application, so the individual
application's HTML forms response time is not really time
lost to the user.  [I do this constantly, opening multiple
windows to the same server or different servers, and
shifting among them as they respond to mine among
potentially millions of other submissions.]

xanthian, UI and GUI designer and implementer since 1976 and
up to about four hours ago, when I got six new windows and
50 individual data entry widgets up in my Java toy after
working on them for the previous 48 hours (you have to love
Java's code reusability). Now I'm putting substance behind
them, or at least will be again as soon as I finish playing
for a while.

[Yes, I just invented and am implementing 38 new heuristics
for Traveller. No, I didn't work on this even half of those
48 hours; most of that time I was out of the house
attending to demands of the real world.]

--
Posted via Mailgate.ORG Server - http://www.Mailgate.ORG

 
 
 

Fat Interfaces (opposing): The Client Is Too Thin

Post by Roedy Gree » Thu, 14 Aug 2003 05:14:00


On Tue, 12 Aug 2003 19:18:27 +0000 (UTC), "Kent Paul Dolan"

Quote:>Yet in the event, user's don't _want_ "beautiful"
>interfaces, they want _minimalist_ _standard_ _already
>learned_ HTML interfaces that add minimal mental effort
>into shifting from application to application.

They also want interfaces that tell them about problems before they
hit submit.

The don't like Java because it takes so long to start and so often the
browser support is flaky and the Applet crashes.

We are offering them a choice of feces or shit.  Because they choose
feces does not mean they would not appreciate a DECENT way to enter
data rapidly.

--
Canadian Mind Products, Roedy Green.
Coaching, problem solving, economical contract programming.
See http://mindprod.com/jgloss/jgloss.html for The Java Glossary.

 
 
 

Fat Interfaces (opposing): The Client Is Too Thin

Post by Nikolaus Maac » Thu, 14 Aug 2003 06:22:19



> Yet in the event, user's don't _want_ "beautiful"
> interfaces, they want _minimalist_ _standard_ _already
> learned_ HTML interfaces that add minimal mental effort
> into shifting from application to application.

This is true.

But I'd also like a complicated and playful interface that opens random
webpages and fills my world with shifting pictures and random music.
The results should be colourful, crazy, and hypnotic.  I'd also like
this program to allow for varying levels of user input -- from none to
lots -- and I want it to run flawlessly on a MAC, pre OSX.

And should any picture of noise catch my interest, I'd like to be able
to click on something that will provide me with the actual URL.

That would be very nice.

     Nik
http://www.nikart.ca

 
 
 

Fat Interfaces (opposing): The Client Is Too Thin

Post by Silvio Bierma » Thu, 14 Aug 2003 07:59:48


I am talking about Web-applications in the sence that we are shifting users
to web-based (read HTML-based) versions of software packages they used to
handle through GUIs (read Windows interfaces).

The initial reaction from users is (and often stays for a long time) that
they preferred the GUI version and experience the HTML version as a decrease
in user-friendlyness.

I agree that the GUI paradigm has lead both programmers and users to create
interfaces that are to clever to be effective and we have overshot our
targets there. HTML has neat limitations that leads us to re-thinking user
interfaces and makes us (finally) come up with simple and sensible
solutions.

It takes a lot of persuasion to make many users see it the same way though.

The HTML interfaces we build using our framework are a magnificent
compromise between HTML simpleness and GUI fancyness. If you are interested
you can check one of the applications. The HTML/JS it uses is only for IE6
(yes I know, it was for good reasons at that time) but using that you might
check http://jambo2.idfix.nl:1880/jambo/manager and log in with username
'gnoodt' and password 'gnoodt' also. It is in dutch and the data will mean
nothing to you, but you must be able to tell the difference with regular
HTML based interfaces. BTW: JS and Cookies are a must here to.

Please notice that browser-dependence is a choice of the application
developer (HTML/JS wise) and has nothing to do with the component framework.

Regards,

Silvio Bierman

 
 
 

Fat Interfaces (opposing): The Client Is Too Thin

Post by Scott Dors » Thu, 14 Aug 2003 10:54:52



Quote:

>We are offering them a choice of feces or shit.  Because they choose
>feces does not mean they would not appreciate a DECENT way to enter
>data rapidly.

Wasn't all of this addressed completely by Project MAC?
--scott

--
"C'est un Nagra.  C'est suisse, et tres, tres precis."

 
 
 

Fat Interfaces (opposing): The Client Is Too Thin

Post by Kent Paul Dola » Fri, 15 Aug 2003 05:17:38




>> Yet in the event, user's don't _want_ "beautiful"
>> interfaces, they want _minimalist_ _standard_ _already
>> learned_ HTML interfaces that add minimal mental effort
>> into shifting from application to application.
> They also want interfaces that tell them about problems
> before they hit submit.

That's not necessarily so, Roedy.  Given modern computer
speeds, a local application, and a split of the web page
forms parsing and serving chores so part of it is on the
local console, response can be effectively instant in the
case of errors.  Build on the fly web pages make it possible
to pop the form back with the errors highlighted and
detailed explanations of what's expected included, without
ever leaving the standard HTML interface from the user's
viewpoint, and still using HTTP to and from the central
server.

This lets users avoid like the plague the "beautiful
interfaces" that are so pretty and so wrongheaded because
non-standard and effortful, making the shift from accounts
payable to project management to human resources to
warehouse services seamless from the viewpoint of learning
another new interface.

Is anybody doing that?  I have no clue about the details,
but that is the effect, at big companies everywhere.

I suspect you have to be working very hard at missing the
point that badly and that foully, Roedy.

I hope whatever is consuming your soul gives you a break
soon.  After cruising your site a fair while back, it has to
be at least as tough being you as it is being me, and I'm
not doing such a great job at the latter task recently.

Quote:> The[y] don't like Java because it takes so long to start
> and so often the browser support is flaky and the Applet
> crashes.

So?  You're part of the software industry.  Volunteer at
Mozilla, and fix that.  Write better Java. Mope and mow
about the JVM not being "load as needed, instant start",
somewhere that it will do some good.

xanthian.

Just finished at dawn, _The Amazing Adventures of Kavalier &
Clay_, a beautiful read you'd probably enjoy, and well
deserving of its Pulitzer.

Now I get to go out and be a witness at the serving of a
subpoena on behalf of my quadrapelegic friend Jesus, if I
can wake John, the process server to be, from his diabetic
stupor.  Myself, it only took me an hour, twice, of coughing
to get myself vertical this morning, one failed about six,
one successful about eleven, and I seem to have been left
running this zoo by default.  Pathetic, but my housekeeping
skills are slowly returning; we may be inside the fire code
in another month.

Work on adding 38 new heuristics and 82 new windows
containing 50 interactive widgets to my Java Genetic
and Memetic Algorithm toy Traveller proceeds apace.

Sojourner is now bigger and much taller than an English
sparrow, and very trusting as I hold it in my clenched*
to feed it, once I get its legs tucked in so it doesn't feel
like it is standing on air.  Mockingbirds are drab, but _so_
cute.

Ark seems to have recovered from all of the diseases he had
at rescue, and I am nearly free of the identifiable ones I
caught from him.

--
Posted via Mailgate.ORG Server - http://www.veryComputer.com/

 
 
 

1. Connectionist Computations based on an Optical Thin-film Model

Hi, there. This year we have conducted several experiments on an alternative
connectionist architecture based on an Optical Thin-Film Model. This is a
multilayer model, with the thicknesses of thin-film layers serving as
adjustable "weights" for the computation. The results are quite encouraging.
Our algorithm can be trained to memorize 3 different patterns, according to
the ideal outputs(reflectance). Their actual presentations are series of real
numbers(reflectance) in terms of different wavelengths. We also gave certain
noise to the trained data, and it was reasonably tolerant. The damaged pattern
can still be recognized.

Is anyone doing similar experiments in this area? I want to share your idea.

And comparisons with other neural network algorithms are also needed from you.

Thanks in advance.

Xiaodong

2. WFW 3.11 MS Mail Problems...

3. NPC AI in a client/server based multiplayer 3d game

4. Mac HD for AppleIIC?

5. Internet Othello Server - how to write a client?

6. Java Console IO

7. Looking for client-server games

8. Looks like "Smoke and Mirrors" from US West.

9. PLEASE RESPOND! CLIENT/SERVER RESEARCH QUESTIONNAIRE

10. eliza in server/client

11. A GAURANTEED WAY TO MAKE MONEY FATS!!!!!!!

12. Win95 OSR 2.1 CD $35 W/Fat32, Fat converter, Full install